The present invention relates to a method of operating an image sensor that has a plurality of light sensor elements. The method includes the following steps for each light sensor element:
a) at an output of the light sensor element, picking up a first signal level representative of a quantity of light picked up by the light sensor element, and
b) resetting the light sensor element and picking up a resulting second signal level.
The present invention also relates to an image sensor that has a plurality of light sensor elements, each having an output for outputting a signal level representative of the quantity of light picked up by the light sensor element, and a reset circuit element for resetting the light sensor elements on the basis of a reset signal that is the same for all light sensor elements.
An image sensor and a method of operating the image sensor are described by S. Decker, R. McGrath, K. Brehmer, and C. Sodini in the 1998, IEEE International Solid State Circuits Conference at pages 176-177.
Resetting the light sensor elements and picking up the resulting second signal level are necessary in order to compensate for irregularities that can occur between individual pixels or light sensor elements of an image sensor, e.g., in the dark current of a photosensitive semiconductor junction or an amplifier, if provided, connected to the semiconductor junction. Such irregularities are combined under the collective term fixed pattern noise (FPN). Compensation is essential for detection of the quantity of light picked up by the light sensor element with a high brightness resolution or high brightness dynamics, in particular when the light sensor elements is designed according to CMOS technology.
Resetting results in the output signal of each light sensor element oscillating between a fixed reset level and a lighting-dependent level at a frequency corresponding to the image pick-up frequency of the image sensor, in which case the difference between the two levels may be substantial. The return from the reset level to the lighting-dependent level is associated with charging or recharging of the capacitance of the light sensor element, which must be accomplished essentially by the photoelectric current of the photosensitive semiconductor. Therefore, the image sensor requires a certain recovery time after each reset before brightness signals can be picked up again by its individual light sensor elements. Although the photoelectric current of a single element can easily be increased by increasing the area of the photosensitive semiconductor junction, this does not permit a higher image pickup frequency, because its capacitance also increases with the area of the junction.
As an additional step in the method according to the present invention, after resetting the light sensor element and before picking up another signal level from the output of the light sensor element, the output is initialized at a signal level that is selected as a function of the first signal level, which was picked up previously from this output and is representative of the light intensity picked up. This additional step is based on the fact that the brightness value picked up by a light sensor element during a given image cycle is usually near the level picked up during the preceding cycle. Consequently, if the output of the light sensor element is initialized at a level which is at least near the level, previously picked up this light sensor element, it shortens the amount of time the light sensor element will need to stabilize at a level representative of the amount of light instantaneously picked up. The signal level selected as a function of the first signal level for initialization is, for example, identical to the first signal level picked up previously.
The effect of this additional step is that the signal level set when the light sensor element is reset can be selected freely and may also be inside the dynamic range of the sensor element. Selecting a signal level outside the dynamic range is supported by the simple circuitry with which this embodiment can be implemented; when a signal level inside the dynamic range is used, a shorter reaction time of the sensor element can be achieved, because operation of the circuit modules in the saturation range is avoided.
This additional step is advantageous in a method where the signal level resulting from the reset is outside the dynamic range of the light sensor element, and thus the difference between the two signal levels picked up is not negligible, regardless of the light intensity picked up by the light sensor element.
In the method and the image sensor according to the present invention, however, the signal level resulting from the reset may also be inside the dynamic range of the sensor element.
This signal level is the same for all image cycles during operation of the image sensor.
An advantageous application of this method is for image sensors where the light sensor element includes a photodiode or a phototransistor operated at a weak inversion. Such image sensors have the advantage that their phototransistors generate a photoelectric current that in turn generates a voltage signal level proportional to the logarithm of the illuminance, and therefore permit a determination of brightness values over a range of up to 8 decades with a performance similar to that of the human eye. This advantage is gained at the expense of extremely low photoelectric currents, which is why the change in signal level associated with the reset is especially problematical with such image sensors.
Corresponding advantages like those achieved with the method according to the present invention are also achieved with the image sensor according to the present invention, which includes an initialization circuit element for initializing the output of a light sensor element at a level that can be preselected for each light sensor element by a control signal received by the initialization circuit element.
Such an image sensor includes, for example, a memory element for storing the signal level representative of the light intensity and for outputting the stored signal level as a control signal to the initialization circuit element. These memory elements are combined, for example, in a common memory array. Combining them in this way permits a space-saving integration of the memory elements. In addition, each memory element can be assigned to different light sensor elements in the time-division multiplex method.
The alternative of integrating a memory element into each individual light sensor element has the advantage that the signal paths are shorter and consequently the sensor element is faster and less sensitive to external interference.
A phototransistor operated at a weak inversion can only supply a signal having a relatively high impedance, so an amplifier is preferably provided for each individual light sensor element. This amplifier is capable of outputting a signal that is relatively insensitive to interference and noise onto a bus line of the image sensor. To minimize the power consumption of the amplifier, it is preferably designed as a voltage follower having unit gain.